Sampling piano - techniques,etc

[Phil]

Hi,
I\'ve clearly got too much time on my hands, so I\'ve decided to sample the steinway concert grand for gs. I\'ve got no problems with recording techniques, but I could really do with some info on what samples to record, how to put them together, etc.
Anybody got any info or sites that can help?
thanks

[esperlad]

Make this piano large. 16 dynamic layers would be great. 8 with the pedal up and 8 with the pedal down. trillls would be nice too.

[Chadwick]

Don\'t forget an option to use release samples for staccato releases.

Have a look at Warren Trachtman\'s site - he has some interesting info on Gigastudio\'s response to various midi velocities, which may affect how you set up your layers once you\'ve finished sampling. I\'m hoping the new patch changes that though.

Hope you\'ve got plenty of hot cocoa for those long tedious recordings. (88 notes lasting 20-30 seconds times about 16 to 20 layers .... eek!)

At least it\'s fairly mechanical compared to sampling wind instruments or choirs etc.,

Good luck

[Duncan Brinsmead]

I would pay attention to getting an accurate and smooth range in the samples at perform time rather than trying to clean it up in the giga patch. Don\'t change the mic position for
different samples and for a given strike layer match the velocities rather than the note volumes( to avoid biasing low or high notes ). I\'m not sure how to accurately perform the different velocities, outside of
attaching a player action( there are pretty good ones now that can be fit to any piano, but they cost a few thousand ). I agree that 16 layers would be great if you can do it.
Sampling the soft pedal might be nice. I don\'t think the lowpass filtering in gigasample is up to getting that soft pedal sound from the samples, but perhaps a second
set of processed samples( perhaps only the pedal down set pp-mf set ) could be created
using some third party software. Another thought is to make a patch where the pp layer is recorded with the soft pedal, although perhaps this would create to harsh a transition. To me the key to a playable instrument is the continuity between keys and color change in the dynamic range. In the final patch the tonal coloration should change with velocity more than the volume. When it is the other way around it sounds like someone is playing with a volume control. I think that you should avoid using any giga filtering altogether.

Personally I wouldn\'t care for any complex samples like trills, as a trill should simply sound the same if the sampling is done well.

Most of the sampled Steinways I\'ve heard are a tad too bright for my tastes. Anything you can do to create a rich warm sound without sounding dull or muffled would be nice.

[Phil]

thanks, been to trachmans site, but couldnt really find anything particularly relevant.
I wasn\'t plannning to do any phrases or trills, but I might look into it at a later date. Likewise, I wasn\'t planning on using any filters. Matching velocities may be difficult, maybe I could sample at eg 20-30 varying volumes, and take some averaged volumes within the maximum range of each note, would that be sensible?
The steinway\'s going to have to be sampled where it is in the concert hall, so I guess I could get a warmer sound from moving the mics slightly further away from the strings, I\'ll experiment until I get a sound that I like and take it from there.
I\'ll let you know how it goes, might be some time.

[Chadwick]

You\'re right Phil, I went back to the site and couldn\'t find the pages either.

Warren mustn\'t have them in his site index, probably because it\'s not what you\'d call \'for general consumption\', but I can assure you the pages are there and are an interesting read.

I might trawl through Warren\'s old posts to these forums. That\'s probably where I found the URLs.